High-Temperature Energy Storage - ACS Publications - American

Subscriber access provided by Fudan University

Article

High temperature energy storage: Kinetic investigations of the CuO/CuO reaction cycle 2

Markus Deutsch, Florian Horvath, Christian Knoll, Daniel Lager, Christian Gierl-Mayer, Peter Weinberger, and Franz Winter Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/acs.energyfuels.6b02343 • Publication Date (Web): 04 Jan 2017 Downloaded from http://pubs.acs.org on January 5, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Energy & Fuels is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

High temperature energy storage: Kinetic investigations of the CuO/Cu2O reaction cycle Markus Deutsch,∗,† Florian Horvath,† Christian Knoll,‡,† Daniel Lager,¶ Christian Gierl-Mayer,§ Peter Weinberger,‡ and Franz Winter† †Institute of Chemical Engineering, TU Wien, Vienna, Austria ‡Institute of Applied Synthetic Chemistry, TU Wien, Vienna, Austria ¶Austrian Institute of Technology (AIT), Vienna, Austria §Institute of Chemical Technologies and Analytics, TU Wien, Vienna, Austria E-mail: [email protected]

Abstract Thermochemical energy storage (TCES) is considered a possibility to enhance the energy utilization efficiency of various processes. One promising field is the application of thermochemical redox systems in combination with concentrated solar power (CSP). There, reactions of metal oxides are in the focus of research, since they allow an increase in process temperature. The reaction system CuO/Cu2 O has been reported as a suitable candidate for TCES. For proper development and modelling of combined CPS-TCES processes, reliable kinetic data are necessary. This work studies the reduction of CuO and the oxidation of Cu2 O under isothermal and isokinetic conditions. The reaction are analysed using a simultaneous thermal analysis (STA) and a lab scale fixed bed reactor. The reaction behaviour shows significant differences between both analyses. To develop kinetic models the non-parametric kinetic (NPK) approach is utilized. This model free approach is expanded by the Arrhenius correlation to increase the applicable

1

ACS Paragon Plus Environment

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 33

temperature range of the models. The resulting models are evaluated and compared. Furthermore, the cycle stability of the system over 20 cycles is assessed for a small sample mass in the STA and a large sample mass in the fixed bed reactor.

Introduction Thermochemical energy storage (TCES) has been in the focus of researchers for the last couple of years due to its advantages and potential compared to other thermal storage technologies. 1 The achieved energy storage densities of TCES-systems are up to five times higher than comparable latent heat storage systems and up to fifteen times higher than sensible heat storage systems. 2 Additionally, the heat is stored in form of chemical energy, which makes energy storage possible without the need for sophisticated and expensive insulation. 3–5 This makes TCES especially suitable for high temperature applications, such as the combination with concentrated solar power (CSP). In this cases, TCES can be used to increase their production during times with low sunlight (e.g. when the sun is blocked by clouds or after sunset). 6,7 Especially, redox systems of metal oxides are deemed to be suitable due to their operating temperature. 8,9 Wong et al. analysed 16 potential metal oxide systems following reaction equation 2. 10 From these systems, only BaO, Co3 O4 , CuO, Fe2 O3 , Mn2 O3 displayed the necessary behaviour for a suitable TCES material. Co3 O4 showed the most promising results, according to Wong, 11 and was further analysed in terms of reversibility and energy storage capacity. 12–14 The cycle ability of Mn2 O3 has already been proven by Carrillo et al. 15 and compared to Co3 O4 . 16

Reduction

Mx Oy → Mx Oy−1 + 1/2O2

∆HR > 0

(1)

Oxidation

Mx Oy−1 + 1/2O2 → Mx Oy

∆HR < 0

(2)

As pointed out by Alonso et al., copper oxide is highly available in emerging CSP mar2


Page 3 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

kets like Mexico and Chile, therefore, reducing the material price compared to other metal oxides. 17 In terms of energy storage capacity CuO is as capable as Co3 O4 (811 kJ/kg for CuO and 844 kJ/kg for Co3 O4 at their theoretical transition temperature in air) and better than Mn2 O3 (202 kJ/kg). 11 Both facts consolidate the potential of CuO as TCES candidate. For proper design of the reactor, in which the solid-state reaction should take place, the knowledge of the reaction kinetics for the applied system is crucial. 18 Additionally, faster reaction kinetics mean that the process has a shorter response time. The kinetics of the system CuO/Cu2 O has been investigated as possible oxygen carrier for chemical looping combustion (CLC). Several groups reported a reduction of the oxidation rate in air at higher temperatures. 19,20 Clayton et al. studied the oxidation and reduction kinetics of Cu-based carrier materials with a CuO loading of up to 50%. 21,22 Their work focused on the development of kinetic expressions aimed at better describing the observed oxidation profiles of cuprous oxide-based oxygen carriers. Zhu et al. investigated the oxidation of Cu2 O in a temperature range of 600-1050 ◦ C during the copper oxidation with and without an initial thin CuO layer. 23 The goal of this work is the development of reliable kinetic models for the reaction system CuO/Cu2 O as it occurs in a TCES application. Previous works investigated the system only for CLC application, in which copper based carriers with CuO loading up to 64% are used. 21,22 For TCES applications would result in an reduced energy density of at least 37%. Therefore, it is important to measure kinetics, without possible effects of carrier material. The kinetic analysis was performed in a simultaneous thermal analysis (STA) and in a lab scale fixed bed reactor. The cycle stability of the system is analysed on a chemical reaction level in an STA and on a macroscopic level in a lab scale fixed bed reactor to further investigate the applicability for TCES.

3


Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 33

Fundamentals CuO/Cu2 O system Used as a TCES material the copper in CuO cycles between the Cu(II) cupric state and the Cu(I) cuprous state

4CuO(s) + ∆HR ⇌ 2Cu2 O(s) + O2 (g)

(3)

For this reaction the equilibrium relation can be derived as a2Cu2 O aO2 K= a4CuO

(4)

where ai represents the activity of the corresponding species. The activity of pure solids (CuO and Cu2 O) can be assumed to be 1, and the activity of O2 can be expressed as its fugacity, with pΘ being the standard pressure of 1 bar.

aO 2 =

f O2 pΘ

(5)

With the simplification that O2 behaves like an ideal gas, then the fugacity is equal to the partial pressure of oxygen pO2 in the reaction atmosphere resulting in

K=

p O2 pΘ

(6)

With the Gibbs equation for free energy

∆GTR = ∆HRT − T ∆SRT = −RT ln(K)

(7)

where ∆GTR , ∆HRT ∆SRT are the free Gibbs energy, the reaction enthalpy and the reaction entropy at temperature T , respectively (6) leads to the stability equation of the system 4


Page 5 of 33

ln(

p O2 −∆HRT ∆SRT ) = + pΘ RT R

(8)

This equation is represented graphically in the stability diagram in Figure 1. As shown there, CuO reacts to Cu2 O in air (pO2 = 0.21 bar) at temperatures over 1028 ◦ C . 24 0

CuO

briu uili

-0.5

e urv mc

eq

0.21 bar

-1

Cu2O

-1.5

2

log(pO )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

isothermal measurements

-2

isokinetic measurements

-2.5 10-5 bar

-5 1028 °C -5.5 900

950

1000 1050 temperature in °C

1100

1150

Figure 1: Stability diagram for CuO/Cu2 O based on data from HSC Database 24 shown with the investigated experimental conditions

Kinetic identification In this work, the non-parametric kinetic (NPK) analysis method was chosen for the kinetic analysis of the CuO/Cu2 O cycle. The NPK method was developed by Serra et al. and Sempere et al. 25–28 for isokinetic measurements and generalized by Heal. 29,30 The method is free of a priori established models and avoids explicit kinetic models as well as the Arrhenius law. In principle the method separates the effect of conversion α and temperature T on the conversion rate dα/dt. In the following the method is described in more detail. Based on the general kinetic equation dα = f (α)k(T ) dt 5


(9)

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 33

the conversion rate dα/dt can be seen as a continuous three dimensional surface based on the conversion dependency f (α) and the temperature dependency k(T ). This surface can be discretized and written in a (n × m) matrix A1 so that the matrix element Ai,j gives the conversion rate for αi and Tj , as shown in equation (10): 

f (α1 )k(T1 ) f (α1 )k(T2 )  f (α2 )k(T1 ) f (α2 )k(T2 )  A= .. ..  . .   f (αn )k(T1 ) f (αn )k(T2 )



f (α1 )k(Tm )  · · · f (α2 )k(Tm )   .. ...  .   · · · f (αn )k(Tm ) ···

(10)

The analysis applied on isothermal experiments results in a matrix A, where each column represents one measurement. The incorporation of isokinetic measurements results in a widely empty matrix, therefore, it requires the implementation of submatrices in A. Further details can be found in literature. 25–27 As a result of the discretization A can be written as matrix product

A = f kT

(11)

with the vectors f and k defined as:

f = f (α1 ) f (α2 ) · · ·

k = k(T1 ) k(T2 ) · · ·

T

(12)

T

(13)

f (αn )

k(Tm )

Utilizing the singular value decomposition (SVD), 31 the matrix A is decomposed into three matrices A = U SV T

(14)

with U and V being orthonormal matrices and S is a diagonal matrix whose entries are 1

bold upper-case letters symbolize matrices, bold lower-case letter vectors

6


Page 7 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

the singular values of A. If the investigated reaction can be described by (9), then only the first entry of S is significantly different from zero. From this follows, that only the first columns of U and V T are significant. Therefore, (14) can be reduced to

A = usv T

(15)

with s = S(1) and u and v T being the first columns of U and V T , respectively. Comparing (15) to (11) then the following correlations can be found:

u = c1 f

(16)

v = c2 k

(17)

1 c1 c2

(18)

s= with c1 and c2 being scaling constants.

Hence, by applying a SVD analysis on A, two independent vectors u and v are found, which are proportional to the conversion dependency f and the temperature dependency k of the reaction. When assuming the temperature dependency of the reaction k(T ) follows the Arrhenius law,

k(T ) = A e−Ea /RT

(19)

then the activation energy Ea and the pre-exponential factor A can be calculated from its linearisation with (17)

ln(k(T )) = ln(vT ) = ln(c2 A) − 7

Ea RT


(20)

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

where vT represents the value of v for temperature T . The pre-exponential factor A cannot be calculated directly from the axis section of (20) since the scales of the quantities are lost in the matrix decomposition process. 29 This can be seen in (18). When c1 is multiplied by an arbitrary factor and c2 is divided by the same factor, the result of (15) would not change, but the vectors u and v would be shifted. Therefore, A can only be calculated if the conversion dependency f (α) is found (this gives c1 ). This can be done by fitting different conversion models onto u. At this point 29 different conversion models for f (α) are known in literature. 32 In this work the identification of Ea has been performed, where possible, since it grants the possibility to increase the temperature range of the kinetic model to a certain extend. The identification of the conversion model does not grant any additional information for the simulation and was, therefore, omitted.

Experimental setup Material R The material used for all experiments was a granular CuO from Merck Emsure . The

particle size distribution of the granulate material can be seen in Figure 2. For further analysis the granulate was ground with a Retsch planetary ball mill PM 100 and afterwards sieved. For the analysis the fraction with a particle size between 32 µm and 45 µm was used. A BET surface analysis resulted in a specific surface of the milled material of 3.1 m2 /g. The phase composition yielding >99,9% pure CuO phase was determined by X-Ray powder diffraction.

Simultaneous thermal analysis - STA The STA measurements where done in a Netzsch STA 449 F1 Jupiter with a differential thermal analysis (DTA) measurement setup. The sample mass in all STA measurements was 151.2 ± 0.6 mg CuO. The measurements were performed according to Figure 3. 8


Page 8 of 33

1

0.5

0.9

0.45

0.8

0.4

0.7

0.35

0.6

0.3

0.5

0.25

0.4

0.2

0.3

0.15

0.2

0.1

0.1

0.05

0

mass fraction

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

cumulative mass fraction

Page 9 of 33

0 0

1

2 3 4 particle size in mm

5

6

Figure 2: Mass based particle size distribution of the CuO granulate. Drawn points represent mean value of mesh size. For isokinetic STA measurements the sample was heated up with a constant heating rate β up to 1100 ◦ C in N2 atmosphere (pO2 < 10−5 bar). Then the atmosphere was switched to synthetic air (pO2 = 0.21 bar) and the sample was cooled down with a cooling rate equal to the heating rate. In this work heating rates of 2, 5 and 10 K/min were applied. For isothermal STA measurements, the sample was heated up under synthetic air atmosphere with 10 K/min up to the measurement temperature. When a constant sample temperature was achieved, the atmosphere was switched to N2 inducing the reduction. After the reduction was completed, the atmosphere was switched back to synthetic air to start the oxidation. At the end of the oxidation the sample was cooled down with 10 K/min. In this work, isothermal STA measurements were performed at 920, 950 and 980 ◦ C. The use of N2 during the reduction was motivated, by the fact that the system CuO/Cu2 O has an eutectic point at 1091◦ C and 35% Cu2 O. 33 A reduction under air (pO2 = 0.21 bar) would require temperatures above 1028◦ C (see Figure 1). Operating the system this close to its melting point could pose problems, like sintering or partial melting, in the operation of the TCES system. The cycle test was performed at 950 ◦ C. The sample was heated up with 10 K/min in synthetic air atmosphere. When isothermal conditions were reached, the cycle was started.

9


N2 0

50

N2+O2

mass temperature

100 99 98 97 96 95 94 93 92 91 90

100 150 200 250 300 350 400 450 500

1100 1000 900 800 700 600 500 400 300 200 100

temperature in °C

1100 1000 900 800 700 600 500 400 300 200 100

Page 10 of 33

mass in %

100 99 98 97 96 95 94 93 92 91 90

mass in %

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

temperature in °C

Energy & Fuels

N2+O2

N2

N2+O2 mass temperature

0

50

time in min

100

150 200 250 time in min

300

350

400

Figure 3: Representative measurement procedure in the STA for kinetic analysis. left: isokinetic measurement with β = 5K/min, right: isothermal measurement with T = 950◦ C The cycle consisted of two segments: 1. reduction under N2 atmosphere for 60 min 2. oxidation under synthetic air atmosphere for 75 min The cycle was repeated 20 times. Then the sample was cooled down with 10 K/min to room temperature. From the mass signal m(t) of each measurement, the mass content of CuO w(t) can be calculated based on

w(t) =

m(t) − m0 ν m0 − m0 ν

with ν =

MCu2 O 2 MCuO

(21)

with MCuO and MCu2 O are the molar masses of CuO and Cu2 O, respectively and m0 the sample mass of CuO at the beginning. The term m0 ν is equal to the theoretical mass of Cu2 O after full conversion. The conversion α in case of the oxidation is equal to w(t), in case of the reduction α = 1 − w(t) is valid.

Fixed bed reactor A reactor was used to investigate the reaction under macroscopic conditions. Its schematic setup is shown in Figure 4. It consists of two tubes on top of each other with an inner 10


Page 11 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

diameter of 40 mm. The lower tube is the preheating zone and is filled with sand for better heat transfer onto the reaction gas. In the reaction chamber the material was placed on a glass frit. The quartz glass frit with a pore size between 160-250 µm is used to evenly distribute the reaction gas over the cross section of the reactor. The reactor is heated with resistance heating shells up to 1100 ◦ C. The temperature control is done based on the temperature inside the reactor, which is measured with a thermocouple type K within the CuO bulk. To achieve different gas atmospheres, N2 and O2 can be mixed via two mass flow controllers. After passing the reactor the reactive gas passes a gas cooling system into a gas analyser which measures the O2 concentration cO2 . Compressed air is used for outer cooling of the reactor and better temperature control. In the reactor isothermal tests at 930, 950 and 980 ◦ C with 50 g of CuO granulate material were performed. The material was heated up with a gas flow of 2.5 l/min N2 /O2 mixture with pO2 = 0.21 bar to simulate air. When isothermal conditions were reached in the reactor the reaction atmosphere was switched to 2.5 l/min N2 (pO2 ≤ 10−5 bar) to induce the reduction (see Figure 1). When the O2 concentration reached zero in the off-gas the reduction was assumed to be completed and the oxidation was started. This was done by switching back to the starting reaction atmosphere of 2.5 l/min N2 /O2 mixture (pO2 = 0.21 bar). The cycle test was similar to the test in the STA. The test used 50.8267 g CuO. The sample was heated up in mixed gas flow of 2 l/min N2 and 0.5 l/min O2 . When isothermal conditions at 950 ◦ C were reached the cycle was started. The cycle consisted of two segments: 1. reduction under 2.5 l/min N2 for 200 min 2. oxidation under 2 l/min N2 and 0.5 l/min O2 for 130 min The cycle was repeated 23 times. The conversion α in this case is calculated on the O2 concentration cO2 following Rt

V˙ ∆cO2 (t)dt 0 α(t) = R tend V˙ ∆cO2 (t)dt 0 11


(22)

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

thermocouple

TRC

Page 12 of 33

gas cooling gas analyzer

CuO/Cu2O glass frit

reaction gas

reaction chamber

AIR

FRC

compressed air for cooling

N2

TR

FRC

FRC

heating shell

O2

preheating zone

reaction gas supply

Figure 4: Schematic setup of the reactor used for kinetic measurements under fixed bed conditions where V˙ is the volumetric gas flow through the reactor, tend the duration of the reaction and ∆cO2 (t) the difference in cO2 between an empty reactor and a reactor filled with reactive material. The change of V˙ due to the released or consumed O2 is negligible.

X-Ray powder diffraction The powder X-ray diffraction measurements were carried out using a PANalytical X’Pert diffractometer in Bragg Brentano geometry using Cu Kα1,2 radiation, an X’Celerator linear detector with a Ni filter, sample spinning with back loading zero background sample holders and 2Θ = 4-90◦ (T = 25 ◦ C). The diffractograms were evaluated using the PANalytical program suite HighScorePlus v3.0d. A background correction and a Kα2 strip were performed.

BET-analyzer The analysis of the specific surface of the samples was determined with nitrogen sorption measurements on an ASAP 2020 (Micromeritics) instrument. The samples (amounts between

12


Page 13 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

100-200 mg) were degassed under vacuum at 80 ◦ C overnight prior to the measurement. The surface area was calculated according to the method of Brunauer, Emmett and Teller (BET). 34

Results STA measurements The results of the isokinetic STA measurements are shown in Figure 5. In the left diagram the reduction of CuO in N2 atmosphere during heating can be seen for different heating rates. It reveals that the system reacts completely regardless of the heating rate with the typical shift of the conversion curve to higher temperatures for higher heating rates. The right diagram shows the oxidation of Cu2 O during cooling in synthetic air atmosphere. There, the reached content of CuO strongly depends on the cooling rate, with the slowest rate reaching the highest CuO content. The reaction stops as the temperature falls below 800 ◦

C. This behaviour is similar for all investigated cooling rates, therefore, it can be concluded

that the available energy at temperature levels below 800 ◦ C is not enough to promote the reaction. Hence the reaction is kinetically limited by the temperature. This also explains the difference in final conversions, as the sample in a measurement with a slower cooling rate remains longer in a temperature range above 800 ◦ C where the reaction occurs. Due to this, the temperatures 920, 950 and 980 ◦ C were chosen for the isothermal measurements, as the reaction has a relevant reaction rate in this temperature range. The isothermal measurements are shown in Figure 6. The left diagram shows the reduction of CuO in N2 atmosphere, the right diagram shows the oxidation of Cu2 O in synthetic air atmosphere. In both cases the conversion is complete. On one hand, it can be seen that higher temperature results in faster reduction. On the other hand the oxidation is slowed down with increased temperature as it has been reported in the literature. 21 Additionally, the isothermal experiments are used to measure the energy turnover during the reaction. Fig13


1

1

0.9

0.9

0.8

0.8

0.7

0.7

mass content CuO

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

mass content CuO

Energy & Fuels

0.6 0.5 0.4 0.3 0.2

0 700

750

800

0.6 0.5 0.4 0.3 0.2

2 K/min 5 K/min 10 K/min

0.1

Page 14 of 33

0.1 850 900 950 temperature in °C

1000

1050

1100

0 700

2 K/min 5 K/min 10 K/min 750

800

850 900 950 temperature in °C

1000

1050

1100

Figure 5: Isokinetic STA measurements at different heating/cooling rates - left: reduction in N2 during heating, right: oxidation in synthetic air during cooling ure 7 shows the DSC signal during a measurement at 950◦ C. The baseline was constructed according to Höhne et al.. 35 The absorbed energy during the reduction (820 kJ/kg CuO) is slightly higher than the released energy during the oxidation (806 kJ/kg CuO). This is most likely due produced oxygen, which exits the cruible during the reduction. Thus taking away some energy, which is measured as additional cooling of the sample. Still, both values are within 1% error to the value reported in the literature (811 kJ/kg CuO 11 )

Reactor measurements The results of the measurements in the test reactor are presented in Figure 8. All measurements reached full conversion, which was verified by X-ray powder diffraction analysis. The major difference between the STA measurements and the reactor measurements is that the oxidation does not slow down with the increasing temperature. This is most likely connected to a better mass transport inside the bulk due to the flow through of the reactive gas compared to the laminar transport in the STA crucible. Additionally, it has to be noted that the temperature inside the reactor for the oxidation is not constant. As a result of the big sample mass in the reactor (50 g), the released energy 14


Page 15 of 33

1

1 920 °C 950 °C 980 °C

0.9 0.8

0.9 0.8 0.7

mass content CuO

0.7 0.6 0.5 0.4 0.3

0.6 0.5 0.4 0.3

0.2

0.2

0.1

0.1

0 10

20

30

40 50 time in min

60

70

80

0 0

90

920 °C 950 °C 980 °C 10

20

30

40 50 time in min

60

70

80

90

Figure 6: Isothermal STA measurements at different temperatures - left: reduction in N2 , right: oxidation in synthetic air

1

1

0.9 0.5

820 kJ/kg

0.7

0

0.6

-806 kJ/kg

0.5 0.4 0.3

N2+O2

N2

-0.5 -1

DSC in mW/mg

0.8

exotherm

0

mass content CuO

mass content CuO

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

N2+O2

0.2

-1.5 DSC signal mass content

0.1 0 0

50

100

150

200 250 time in min

300

350

-2 400

Figure 7: Energy uptake / release during the conversion in an isothermal measurement at 950◦ C

15


Energy & Fuels

influences the local sample temperature inside the reactor. Due to the thermal inertia of the system, the released energy results in non-isothermal measurement conditions, which are recorded by the thermocouple within the bulk material. Figure 9 shows representatively the deviation of the isothermal temperature of 950 ◦ C during the oxidation. The released energy can hardly be compensated by the cooling of the reactor. This further consolidates the system CuO/Cu2 O as a promising CSP-TCES system. 1

1 930 °C 950 °C 980 °C

0.9 0.8

0.9 0.8 0.7 mass content CuO

0.7 mass content CuO

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 33

0.6 0.5 0.4 0.3

0.6 0.5 0.4 0.3

0.2

0.2

0.1

0.1

0 0

50

100 150 time in min

200

930 °C 950 °C 980 °C

0 0

10

20

30

40 50 time in min

60

70

80

90

Figure 8: Measurements in the fixed bed reactor at different starting temperatures - left: reduction in N2 , right: oxidation in synthetic air

Kinetic identification All models presented in this work can be downloaded at the projects homepage solidheat.project.tuwien.ac.at Reduction The kinetic identification of the reduction was performed on the isokinetic STA measurements (Figure 5). Utilizing the NPK method, this resulted in a model based on (15), with s = 2.949 · 10−3 s−1 and the vectors u and v as shown in Figures 10 and 11. Linearisation of v based on (20) results in the Arrhenius plot in Figure 12. From the slope of the fitted linear 16


Page 17 of 33

20

0.9

17.5

0.8

15

0.7

12.5

0.6

10 mass content temperature

0.5

7.5

0.4

5

0.3

2.5

0.2

0

0.1

-2.5

0

temperature difference in °C

mass content CuO

1

-5 0

10

20

30 40 time in min

50

60

Figure 9: Temperature change in the reaction zone due to the released energy during oxidation at 950 ◦ C starting temperature equation the apparent activation energy Ea for the reduction can be calculated as 255.68 kJ/mol. 1 0.9 0.8 0.7 u - normalized

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

0.6 0.5 0.4 0.3 0.2 0.1 0 0

0.2

0.4 0.6 conversion α

0.8

1

Figure 10: Conversion dependency vector u derived from isokinetic measurements of the reduction of CuO in N2 Therefore, both the isokinetic measurements as well as the isothermal measurements were simulated based on the model derived from the isokinetic data. Additionally, the simulation was performed in two different ways. With the kinetic information calculated with the NPK method, the reduction was simu17


Energy & Fuels

1 0.9 0.8

v - normalized

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 850

950 1000 temperature in °C

900

1050

Figure 11: Temperature dependency vector v derived from isokinetic measurements of the reduction of CuO in N2

0 -0.5 -1 ln(v) - normalized

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 33

-1.5 -2

y = -3.075∙104 x + 23.23

-2.5 -3 -3.5 -4 -4.5 0.75

0.8

0.85

0.9

1000/T in 1/K

Figure 12: Arrhenius plot of v derived from isokinetic measurements of the reduction of CuO in N2

18


Page 19 of 33

lated in two ways. Once, with the result of the NPK method used directly to simulate the conversion (’full NPK’) and once with the temperature dependency described by the identified Arrhenius equation (’NPK + Arrhenius’). Often the conversion dependency found by the NPK method is valid in a sufficient range, but the temperature dependency is limited. By fitting the Arrhenius equation an extrapolation is possible and therefore the temperature range in which the model is applicable is increased. Note that the apparent pre-exponential factor is only correct in combination with the vector u, identified by the NPK method (see chapter ). Figure 13 shows the fit of the simulated isokinetic measurements to the actual measurements for conversion and conversion rate. Both methods reproduce the measurements well and while the full NPK method is limited to temperature above 850 ◦ C due to u (Figure 11), the combination of NPK and Arrhenius equation is capable of giving the behaviour of the reaction below 850 ◦ C. x 10-3 3

1 2 K/min 5 K/min 10 K/min full NPK NPK + Arrhenius

0.8 0.7

2 K/min 5 K/min 10 K/min full NPK NPK + Arrhenius

2.5

conversion rate dα/dt

0.9

conversion α

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

0.6 0.5 0.4 0.3 0.2

2 1.5 1 0.5

0.1 0 700

750

800


1000

1050

1100

0 700

750

800


1000

1050

1100

Figure 13: Simulation of the isokinetic measurements of the reduction of CuO with the identified model The results of the simulation of the isothermal experiments are shown in Figure 14. There it can be seen that the models fit the measurements reasonably well, especially since the model was identified from isokinetic data. Additionally, the reduction in the reactor is analysed. Since only three temperature 19


Energy & Fuels

x 10-3 10

1

920 °C 950 °C 980 °C full NPK NPK + Arrhenius

0.9 8 conversion rate dα/dt


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 33

0.6 0.5 0.4 920 °C 950 °C 980 °C full NPK NPK + Arrhenius

0.3 0.2 0.1

6

4

2

0

0 0

10

20

30

40 50 time in min

60

70

80

90

0

10

20

30

40 50 time in min

60

70

80

90

Figure 14: Simulation of the isothermal measurements of the reduction of CuO with the model identified from isokinetic measurements levels are available, vector v consists of only 3 points. The activation energy Ea based on the corresponding Arrhenius plot (Figure 15) was calculated as 316.9 kJ/mol. The difference between Ea in the STA and the fixed bed reactor leads to the conclusion that the reaction in the STA is not only govern by the intrinsic reaction kinetic but also by mass transfer. 36 One can also see that the conversion dependency u (Figure 16) differs from the dependency found in the STA measurements. Note, the dead volume in the reactor smears the change of the O2 concentration in the off gas. This is especially significant at the beginning of the reduction, when the atmosphere in the reactor is changed from synthetic air to N2 . This causes the steep drop at the beginning of u. Thus, it is assumed that the conversion dependency is only valid for α > 0.1. With the identified model for the reactor, it was also possible to reproduce the measurement results satisfactory (Figure 17). Oxidation Since the isokinetic STA measurements of the oxidation reaction (see Figure 5) do not reach full conversion, it is not possible to calculate the conversion dependency for a satisfying conversion range. Therefore, the NPK method was performed on the isothermal STA ex20


Page 21 of 33

0

ln(v) - normalized

-0.2 -0.4 -0.6

y = -3.812∙104 x + 30.43

-0.8 -1 -1.2 -1.4 0.795

0.80

0.805

0.81 0.815 0.82 1000/T in 1/K

0.825

0.83

0.835

Figure 15: Arrhenius plot of v derived from isothermal measurements of the reduction of CuO in the reactor

1 0.9 0.8 0.7 u - normalized

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

0.6 0.5 0.4 0.3 0.2 0.1 0 0

0.2


0.8

1

Figure 16: Conversion dependency vector u derived from isothermal measurements of the reduction of CuO in the reactor

21


1

0.9

0.9

0.8

0.8

0.7

0.7

0.6 0.5 0.4 0.3


0.2 0.1 0 0

50

100 150 time in min

200

conversion rate dα/dt

1

Page 22 of 33


0.6 0.5 0.4 0.3 0.2 0.1 0 0

50

100 150 time in min

200

Figure 17: Simulation of the isothermal measurements of the reduction of CuO in the testrig periments. The identified vectors u and v are shown in Figures 19 and 18. The shape of u in the oxidation differs from the reduction, thus, it can be concluded that the conversion mechanism differs between reduction and oxidation. Again, vector v consists only of three points. Also, as expected from the isothermal measurement data, the fitted linear equation in the Arrhenius plot of v has a positive slope, resulting in a apparent activation energy Ea of -137.10 kJ/mol. Nevertheless, the simulation can reproduce the measurement with good accuracy (Figure 20). 0 -0.1 -0.2 ln(v) - normalized

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

conversion α

Energy & Fuels

y = 1.649∙104 x - 13.8 -0.3 -0.4 -0.5 -0.6 -0.7 0.80 0.805 0.81 0.815 0.82 0.825 0.83 0.835 0.84 1000/T in 1/K

Figure 18: Arrhenius plot of v derived from isothermal STA measurements of the oxidation of Cu2 O

22


Page 23 of 33

1 0.9 0.8

u - normalized

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

0.2


0.8

1

Figure 19: Conversion dependency vector u derived from isothermal STA measurements of the oxidation of Cu2 O

x 10-4 20

1

18

0.8

16 conversion rate dα/dt

0.9

0.7 conversion α

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

0.6 0.5 0.4 920 °C 950 °C 980 °C full NPK NPK + Arrhenius

0.3 0.2 0.1


14 12 10 8 6 4 2

0

0 0

10

20

30

40 50 time in min

60

70

80

90

0

10

20

30

40 50 time in min

60

70

80

90

Figure 20: Simulation of the isothermal STA measurements of the oxidation of Cu2 O

23


Energy & Fuels

The result of the kinetic identification of the oxidation in the reactor is shown in Figures 21 and 22. In contrast to the other isothermal measurements, vector v consists of more than 3 points. This is due to the non-isothermal condition described in chapter . This can be considered in the kinetic identification by the NPK method and therefore additional temperature information is extracted from the data. As a downside, a non-isotropic temperature field occurs, reducing the quality of the identified model and leading to a higher deviation of the simulation from the measurement (Figure 17). 0 -0.2 -0.4 ln(v) - normalized

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 33

-0.6 -0.8 -1 -1.2 -1.4 0.795

0.80

0.805

0.81 0.815 0.82 1000/T in 1/K

0.825

0.83

0.835

Figure 21: Arrhenius plot of v derived from isothermal measurements of the oxidation of Cu2 O in the reactor

Cycle test The cycle stability test in the STA is shown in Figure 24. The diagram shows the change of the conversion in the sample and the energy flow over 20 cycles. One can see that the form of the conversion curve changes only minor over the cycles. There is no sign of chemical degradation, in fact after 10 cycles the system reaches nearly theoretical conversion. An equivalent cycle test was performed in the reactor, with the result that the granular material sintered together, blocking the reaction after a few cycles. Figure 25 shows the sintering of the material after the first and the last cycle. Additionally, the sintered material 24


Page 25 of 33

1 0.9 0.8

u - normalized

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

0.2


0.8

1

Figure 22: Conversion dependency vector u derived from isothermal measurements of the oxidation of Cu2 O in the reactor

1 0.9 0.8 0.7 conversion α

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

0.6 0.5 0.4 0.3 930 °C 950 °C 980 °C full NPK

0.2 0.1 0 0

10

20

30

40 50 time in min

60

70

80

90

Figure 23: Simulation of the isothermal measurements of the oxidation of Cu2 O in the reactor

25


Energy & Fuels

time in h 1

0

5

10

15

20

30

25

35

40

45

0.9 0.8 0.7 mass content CuO

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 33

0.6 0.5 0.4 0.3 0.2 0.1 0 1

2

3

4

5

6

7

8

9

10 11 12 cycle number

13

14

15

16

17

18

19

20

Figure 24: CuO reaction behaviour over 20 cycles in the STA at 950 ◦ C and changing N2 /synthetic air atmosphere

26


Page 27 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

was cut in half (Figure 25D) . It can be seen that the surface of the material sintered completely together, while the granular structure is still present in the inner core of the sample. Thus, blocking most of the reactive surface. Additionally, cavities of Cu2 O were formed in the edge regions, which was confirmed by a XRD analysis. Based on this behaviour, it can be concluded that a fixed bed process is not practical for energy storage. An alternative option would be to use a rotary reactor, as it prevents the sintering due to relative movement of the particles, as suggested by Alonso et al.. 17 A

C

B

D Cu2O

Figure 25: Sintering of the granulate material in the fixed bed reactor - A & B: top and side view after 1 cycle, C: top view after 20 cycles, D: inside view of the material after 20 cycles

Conclusion In this work the kinetic behaviour and the cycle stability of the TCES system CuO/Cu2 O were analysed in a STA and a fixed bed reactor. Based on isokinetic and isothermal mea27


Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

surements kinetic models were derived for oxidation and reduction. To identify these model the NPK method was applied, which is an a priori model free approach. Then the Arrhenius correlation was introduced to the models to expand the applicable temperature range. The measurements revealed a distinctively different behaviour between the oxidation in the STA and the fixed bed reactor. In the STA, the oxidation slowed down with increased temperature as described in the literature. This results in a negative apparent activation energy Ea of -137.10 kJ/mol. Such behaviour was not observed in the fixed bed reactor, where a mechanistic change was detected, through a change of Ea , but no negative Ea has been found. A comparison between the available literature 21,22,37 and this work shows significant differences in the kinetic of the reaction system. Clayton et al. presented conversion times around 1 to 2 min for the reduction, Hu et al. reported the reaction occur within 10 min while the reaction in this work took around 40 min for full conversion. All three publications conducted their measurements in a TGA or STA. The main difference between all publication is the sample material. While Clayton et al. used copper based oxygen carriers with TiO2 and ZrO2 as support material, Hu et al. used an oxygen carrier based Al2 O3 . The support material reduces the CuO content to around 40%, which directly reduced the energy density. Thus, this work used granular CuO. Johansson et al. reported that different support materials have different effects on the reactivity of system Mn3 O4/MnO. A similar effect could explain the different kinetic behaviour. Further examinations are needed to identify the decisive factor behind this different behaviour. In the fixed bed reactor, the system showed a high energy output during oxidation, proofing the capability of CuO as a TCES material. In the STA the reaction system shows a excellent cycle stability over the course of 20 cycles. In the fixed bed reactor, heavy sintering occurred which reduced the reaction to a minimum by blocking most of the reactive surface. This allowed only the outmost layer of the material to react. Further investigations shows that the sample in the STA also sintered, but due to the small thickness of the sample in the crucible, no reaction hindering effect was measured.

28


Page 28 of 33

Page 29 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Thus, a fixed bed process is not suitable for application, instead a system with movement of the material to prevent sintering should be considered. Overall, the study shows, that the system CuO/Cu2 O is a valid candidate for TCES in combination with CSP, but further investigations are necessary.

Acknowledgement The authors thank the Austrian Research promotion Agency (FFG) for their financial support of the project SolidHeat Kinetics (#848876). The X-ray center (XRC) of the TU Wien is acknowledged for providing access to the powder X-ray diffractometers.

References (1) Aydin, D.; Casey, S. P.; Riffat, S. The latest advancements on thermochemical heat storage systems. Renewable and Sustainable Energy Reviews 2015, 41, 356–367. (2) Abedin, A. H.; Rosen, M. A. A critical review of thermochemical energy storage systems. Open Renewable Energy J 2011, 4, 42–46. (3) Nagel, T.; Shao, H.; Singh, A. K.; Watanabe, N.; Roßkopf, C.; Linder, M.; Wörner, A.; Kolditz, O. Non-equilibrium thermochemical heat storage in porous media: Part 1 – Conceptual model. Energy 2013, 60, 254–270. (4) Li, T.; Wang, R.; Kiplagat, J. K.; Kang, Y. Performance analysis of an integrated energy storage and energy upgrade thermochemical solid–gas sorption system for seasonal storage of solar thermal energy. Energy 2013, 50, 454–467. (5) Michel, B.; Mazet, N.; Mauran, S.; Stitou, D.; Xu, J. Thermochemical process for seasonal storage of solar energy: Characterization and modeling of a high density reactive bed. Asia-Pacific Forum on Renewable Energy 2011 2012, 47, 553–563. 29


Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(6) Romero, M.; Steinfeld, A. Concentrating solar thermal power and thermochemical fuels. Energy Environ. Sci. 2012, 5, 9234–9245. (7) Zaversky, F.; García-Barberena, J.; Sánchez, M.; Astrain, D. Transient molten salt twotank thermal storage modeling for CSP performance simulations. Solar Energy 2013, 93, 294–311. (8) Deutsch, M.; Müller, D.; Aumeyr, C.; Jordan, C.; Gierl-Mayer, C.; Weinberger, P.; Winter, F.; Werner, A. Systematic search algorithm for potential thermochemical energy storage systems. Applied Energy 2016, 183, 113–120. (9) Pardo, P.; Anxionnaz-Minvielle, Z.; Rougé, S.; Cognet, P.; Cabassud, M. Ca(OH)2/CaO reversible reaction in a fluidized bed reactor for thermochemical heat storage. Solar Energy 2014, 107, 605–616. (10) Wong, B.; Brown, L.; Schaube, F.; Tamme, R.; Sattler, C. Oxide based thermochemical heat storage. Proceedings of the 16th SolarPACES Conference, Perpignan, France. 2010; pp 21–24. (11) Wong, B. Thermochemical heat storage for concentrated solar power: Thermochemical system reactor design for thermal energy storage: Phase II Final Report for the period September 30, 2008 through April 30, 2011. 2011. (12) Neises, M.; Tescari, S.; Oliveira, L. d.; Roeb, M.; Sattler, C.; Wong, B. Solar-heated rotary kiln for thermochemical energy storage. Solar Energy 2012, 86, 3040–3048. (13) Block, T.; Knoblauch, N.; Schmücker, M. The cobalt-oxide/iron-oxide binary system for use as high temperature thermochemical energy storage material. Thermochimica Acta 2014, 577, 25–32. (14) Agrafiotis, C.; Roeb, M.; Schmücker, M.; Sattler, C. Exploitation of thermochemical

30


Page 30 of 33

Page 31 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

cycles based on solid oxide redox systems for thermochemical storage of solar heat. Part 1: Testing of cobalt oxide-based powders. Solar Energy 2014, 102, 189–211. (15) Carrillo, A. J.; Serrano, D. P.; Pizarro, P.; Coronado, J. M. Thermochemical heat storage based on the Mn2O3/Mn3O4 redox couple: influence of the initial particle size on the morphological evolution and cyclability. J. Mater. Chem. A 2014, 2, 19435– 19443. (16) Carrillo, A. J.; Moya, J.; Bayón, A.; Jana, P.; La Peña O’Shea, V. A. d.; Romero, M.; Gonzalez-Aguilar, J.; Serrano, D. P.; Pizarro, P.; Coronado, J. M. Thermochemical energy storage at high temperature via redox cycles of Mn and Co oxides: Pure oxides versus mixed ones. Solar Energy Materials and Solar Cells 2014, 123, 47–57. (17) Alonso, E.; Pérez-Rábago, C.; Licurgo, J.; Fuentealba, E.; Estrada, C. A. First experimental studies of solar redox reactions of copper oxides for thermochemical energy storage. Solar Energy 2015, 115, 297–305. (18) Solé, A.; Fontanet, X.; Barreneche, C.; Martorell, I.; Fernández, A. I.; Cabeza, L. F. Parameters to take into account when developing a new thermochemical energy storage system. 1st International Conference on Solar Heating and Coolingfor Buildings and Industry (SHC 2012) 2012, 30, 380–387. (19) Abad, A.; Mattisson, T.; Lyngfelt, A.; Rydén, M. Chemical-looping combustion in a 300 W continuously operating reactor system using a manganese-based oxygen carrier. Fuel 2006, 85, 1174–1185. (20) Eyring, E. M.; Konya, G.; Lighty, J. S.; Sahir, A. H.; Sarofim, A. F.; Whitty, K. Chemical Looping with Copper Oxide as Carrier and Coal as Fuel. Oil Gas Sci. Technol. – Rev. IFP Energies nouvelles 2011, 66, 209–221. (21) Clayton, C. K.; Sohn, H. Y.; Whitty, K. J. Oxidation Kinetics of Cu2O in Oxygen

31


Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Carriers for Chemical Looping with Oxygen Uncoupling. Ind. Eng. Chem. Res. 2014, 53, 2976–2986. (22) Clayton, C. K.; Whitty, K. J. Measurement and modeling of decomposition kinetics for copper oxide-based chemical looping with oxygen uncoupling. Applied Energy 2014, 116, 416–423. (23) Zhu, Y.; Mimura, K.; Isshiki, M. Oxidation Mechanism of Cu2O to CuO at 600–1050◦ C. Oxidation of Metals 2004, 62, 207–222. (24) Antti Roine, HSC Chemistry 6.12, Outotec Research Oy. 2007. (25) Serra, R.; Sempere, J.; Nomen, R. A new method for the kinetic study of thermoanalytical data:: The non-parametric kinetics method. Thermochimica Acta 1998, 316, 37–45. (26) Sempere, J.; Nomen, R.; Serra, R. Progress in Non-parametric Kinetics. Journal of Thermal Analysis and Calorimetry 1999, 56, 843–849. (27) Serra, R.; Nomen, R.; Sempere, J. The Non-Parametric Kinetics A New Method for the Kinetic Study of Thermoanalytical Data. Journal of Thermal Analysis and Calorimetry 1998, 52, 933–943. (28) Sempere, J.; Nomen, R.; Serra, R.; Soravilla, J. The NPK method: An innovative approach for kinetic analysis of data from thermal analysis and calorimetry. Thermochimica Acta 2002, 388, 407–414. (29) Heal, G. R. A generalisation of the non-parametric, NPK (SVD) kinetic analysis method: Part 1. Isothermal experiments. Thermochimica Acta 2005, 426, 15–21. (30) Heal, G. R. A generalisation of the non-parametric, NPK (SVD) kinetic analysis method: Part 2. Non-isothermal experiments. Thermochimica Acta 2005, 426, 23–31.

32


Page 32 of 33

Page 33 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

(31) Golub, G. H.; Reinsch, C. Singular value decomposition and least squares solutions. Numerische Mathematik 1970, 14, 403–420. (32) Dickinson, C. F.; Heal, G. R. Solid–liquid diffusion controlled rate equations. Thermochimica Acta 1999, 340–341, 89–103. (33) Neumann, J. P.; Zhong, T.; Chang, Y. A. The Cu−O (Copper-Oxygen) system. Bulletin of Alloy Phase Diagrams 1984, 5, 136–140. (34) Brunauer, S.; Emmett, P. H.; Teller, E. Adsorption of Gases in Multimolecular Layers. J. Am. Chem. Soc. 1938, 60, 309–319. (35) Höhne, G.; Hemminger, W. F.; Flammersheim, H. J. Differential Scanning Calorimetry; Springer Berlin Heidelberg, 2013. (36) Szekely, J. Gas-solid reactions; Elsevier, 2012. (37) Hu, W.; Donat, F.; Scott, S. A.; Dennis, J. S. Kinetics of oxygen uncoupling of a copper based oxygen carrier. Applied Energy 2016, 161, 92–100. (38) Johansson, M.; Mattisson, T.; Lyngfelt, A. Investigation of Mn3O4 With Stabilized ZrO2 for Chemical-Looping Combustion. Carbon Capture and Storage 2006, 84, 807– 818.

33


High-Temperature Energy Storage - ACS Publications - American

Recommend Documents